Entity risk scoring prerequisites

edit

To use entity risk scoring, your role must have certain cluster, index, and Kibana privileges. This feature requires a Platinum subscription or higher.

This page covers the requirements and guidelines for using the entity risk scoring feature, as well as its known limitations.

Privileges

edit

To turn on the risk scoring engine, you need the following privileges:

Cluster Index Kibana
  • manage_index_templates
  • manage_transform

all privilege for risk-score.risk-score-*

Read for the Security feature

Elasticsearch resource guidelines

edit

Follow these guidelines to ensure clusters have adequate memory to handle data volume:

  • With 2GB of Java Virtual Machine (JVM) heap memory, the risk scoring engine can safely process around 44 million documents, or 30 days of risk data with an ingest rate of 1000 documents per minute.
  • With 1GB of JVM heap, the risk scoring engine can safely process around 20 million documents, or 30 days of risk data with an ingest rate of around 450 documents per minute.

Known limitations

edit
  • You can only enable the risk scoring engine in a single Kibana space within a cluster.
  • The risk scoring engine uses an internal user role to score all hosts and users, and doesn’t respect privileges applied to custom users or roles. After you turn on the risk scoring engine for a Kibana space, all alerts in the space will contribute to host and user risk scores.